Improve readme #119
Improve readme #119
Conversation
README.md
Outdated
|
|
||
| Resalloc's concept of _resources_ is intentionally vague and general. Whatever | ||
| your allocation scripts spawn and return to resalloc, those are your | ||
| _resources_. It can be virtual machines, docker containers, IP addresses, TODO, |
There was a problem hiding this comment.
I wanted to emphasize that resalloc doesn't necessarily have to be used only for VM spawning, so I wanted to list a couple of different resources. Do you have any good ones?
There was a problem hiding this comment.
Whatever you put here; disk volumes, network devices, ... I think it is fine eventually. It could work even with people :-) even though the abstraction would deserve yet another iteration to better support asynchronous "human resources".
| self-hosted sever is for free. Resalloc prioritizes to be wallet-friendly. | ||
| - Web interface - It is possible to optionally run a web UI and let users see | ||
| what, and how many resources are available. For example, see | ||
| [Copr resources](https://download.copr.fedorainfracloud.org/resalloc). |
There was a problem hiding this comment.
The list is long enough ... but did I forget something important?
There was a problem hiding this comment.
This is awesome, thank you very much for the docs! We can continue in a different PR, but something that comes to my mind:
- resource recycling and security -> I can get the very same resource again for multiple tasks just to not waste the resource allocation time, and potentially fixed price for spawning
- fair client service -> despite the overall complicated logic in allocation, the sooner the client comes the earlier is served
- logging -> each resource has it's own set of "live" logs you may monitor by
tail -f
README.md
Outdated
| Typical client use-cases | ||
| ------------------------ | ||
|
|
||
| 1. get the resource, and wait till it is ready |
There was a problem hiding this comment.
this is old text, but "ask for a resource" would sound better
README.md
Outdated
| small number of resources that are ready to be used immediately. | ||
| - Livechecks - Clouds are unreliable. VMs can break while starting or become | ||
| unresponsive for various reasons. Resalloc periodically checks the livelines | ||
| of all resources and makes sure we doesn't leak money. |
There was a problem hiding this comment.
"sure we doesn't leak money" => "makes sure money doesn't leak?"
|
Updated |
|
Thank you! |
| [Copr resources](https://download.copr.fedorainfracloud.org/resalloc). | ||
|
|
||
| Resalloc was created to accomodate the ever growing | ||
| [farm of Copr builders](https://pavel.raiskup.cz/blog/copr-farm-of-builders.html). |
There was a problem hiding this comment.
To be even more precise, the history is much longer - I used to have a shell-only solution (I don't think it was ever published). The motivation was to have the same VM allocator tool both for Copr (the Red Hat) instance and the CI (Jenkins based) for the https://github.com/sclorg images (the team doesn't use resalloc anymore).
I am converting the README to markdown, adding the list of known resalloc allocation scripts, and motivation for why to use the project.